Impacts of Consumer Demands on CATV Local-Loop Communications. Terrence P. McGarty and Sara J. McGarty. Warner Amex Cable Communications Inc

Transcription

1

2 f -'- t I / I! Impacts of Consumer Demands on CATV Local-Loop Communications Terrence P. McGarty and Sara J. McGarty Warner Amex Cable Communications Inc. ABSTRACT There is a growing consumer demand for various services that are based on electronic home delivery. The most viable delivery medium appears to be CATV-cable networks which currently reach 50% of U.S. households. This paper addresses the services that will be provided and presents system architectures to implement these. Various technological choices are available, and based upon a cost benefit study, the technologies selected are dramatically different than those in pure commer~ial networks. This paper concludes with an assessment of key trends in this area. INTRODUCTION Cable networks are expanding at a dramatic rate in large metropolitan areas where population densities are high and consumer demographics are highly favorable to the introduction of many new services. Cable, or CATV as it was originally termed, was a one-way distribution link located in highly rural areas. The new cable systems are in such cities as Dallas, Houston, Pittsburgh, San Diego, Tucson and New Orleans. These systems serve 350,000 to 750,000 households within excess of 1,000 miles of interconnected plant. This plant can support up to 800 MHz of bandwidth with in excess of 100 video channels and having a twoway capability activated in most cases. The cable companies, called MSO's (Multiple Systems Operators), have generally focused separately on the consumer and commercial markets. The consumer uses a cable called the "subscriber cable" that is connected to a cable headend via trunk distribution cables. These cables are MHz in bandwidth with 50 MHz upstream from the subscriber to the head end providing a two-way capability. The subscriber cable is also divided into 6 MHz channels consistent with standard TV channel allocations. The institutional cable usually shadows the subscriber cable and is equally divided in bandwidth in upstream and downstream. These cables were introduced to support local communications for local government. MSO's also envisioned using these for the purpose of commercial communications. This market, however, has been developing slowly. In most cable companies, there is a growing interest in configuring the institutional cable to support standard data services using packet technology or dedicated data circuits. The most dramatic growth areas appear to be on the consumer. side with the provision of a wide variety of services. The reasons for this dramatic growth are as follows: The subscriber cable has rapid penetration into 50% of U.S. households. Suppliers of new services see this as a highly attractive approach to direct merchandising. Consumers are demanding more control and convenience in their purchases. There are low cost technological alternatives that make this new form of a marketing channel attractive as a financial investment. The growth of two-way consumer systems started in 1977 with the Warner Cable QUBE system which provided for a two-way interactive system using a microprocessor-based home keyboard terminal and a polling based multiple access design. More recently other cable companies such as Cox have attempted to develop such systems but at present there are no other operational systems. The Warner system in Pittsburgh will soon have close to 150,000 subscribers of whom 80,000 will be QUBE interactive subscribers. This represents the largest computer time-sharing operation in the world. Viewed in this fashion, the consumer cable services represent one of the greatest technical and operational challenges in distributed systems. The new services that are now being developed in two-way cable systems are: Banking and Financial Services Games and Entertainment Merchandise Shopping Information and Education Classified and Local Services Electronic Mail Travel and Ticketing CH1874-7/83/ $ IEEE 1025,.

3 All of these services will be provided directly to the home using an interactive home terminal (home computer), the cable system, and a central computer complex. This paper discusses a configuration providing these services using this new medium. It discusses needs and demands, system requirements, architecture alternatives, technology and trends. The objective of this paper is to show that the growth of local-loop communications in cable will be dominated both in technology and policy by the consumer side rather than the commercial. It will be this early consumer dominance that will introduce a bias that in turn will drastically shape the future of commercial services. NEEDS AND DEMANDS Too often, technologists develop technologies in search of a market. The classic case is now the use of videotex. This was a technology developed in England by the British Post Office (BPO) with the intent to increase phone usage. It was a response not to an un-met consumer need but to an underfulfilled BPO desire to increase revenue. The result is that after five years there are only 2,000 home subscribers, most of whom are travel agents. Thus, it is essential to first understand the needs and demands of the market and to provide a techno~ogical solution to them and not the opposite. On the consumer side, there is an increased demand for choice and convenience, for courtesy and quality in the delivery of goods and services. There is also a growing acceptance of direct merchandising to the home and a willingness to interact with sophisticated electronic devices (bank ATM's, Atari games, electronic data bases such as CompuServe, and personal computers such as Timex Sinclair). The un-met need is to shop and obtain services at the convenience of the consumer in a frienuly and entertaining manner and to also avoid the inconvenience of in-store shopping with crowds and delays. The supplier of these goods and services also has needs that can be met: a way to expand markets without building new stores and hiring new staff, a means of directly connecting purchases to existing computer systems to reduce operating expenses and inventory costs. Such stores as Sears are already in the position of having IBM Series I in most stores connected to their large CPU's and in-store terminals. The next natural extension is to the home. To summarize, the system, to match consumer and supplier needs, must have the following general parameters: Readily access a large consumer base. Possess a "user-friendly" software to allow frequency of use. SYSTEM Provide a capability to gateway to larger CPU's and allow significant real time transaction capability. Allow many users (100,000 to 500,000) to utilize the systems with a peak instantaneo~ number in the range of 1,000 to 5,000. Provide a real time distributed data base. Provide access to the peak instantaneous number using a multiple access scheme that is cost effective and efficient. REQUIREMENTS There are three main interfaces that the consumer can have to the system: Voice - to assist in providing details too cumbersome to present through a computeronly interface. Data - this is a text only mode. The data provides for the videotex screens with the two-way interactive capability. The text capability provides informational capability. The interface is through a keyboard or keypad. Video - since cable can provide full motion video, and this can be provided on a demand basis, the consumer must have the capability to view video segments on demand. Video has the capability to promote and persuade product. The classic videotex systems provide only data text that informs. The economics of such a system do not indicate a viable business. On the other hand, video has the promotional and persuasive capability to induce transactions. It is these transactions that result in a substantial revenue flow to support a viable business. As a result of this analysis, the system would function in the following manner. The consumer would turn the system on and automatically have a session established and view a text screen. The consumer would readily move through the text data base by responding on a keyboard and having that response transmitted and the desired text screen sent to the consumer and displayed. When the consumer seeks detailed descriptions, a response is entered in their keyboard and a full motion video segment is then transmitted and displayed to the consumer. The segment should be long enough to provide the consumer with adequate understanding of the product (30 to 270 seconds). When the consumer indicates that a transaction occurs, then this data is entered by the consumer in the keyboard, transmitted to the headend computer, and sent via a gateway to the supplier's computer system

4 The supplier's computer system then responds and provides the consumer with a completed transaction verification. This operational scenario responds to the consumer and supplier needs. It represents a technology that meets a market need. ARCHITECTURE The system architecture that meets these requirements is shown in Figure 1. The main elements of the architecture are as follows: Home Terminal - a keyboard microprocessor based Interface device that uses the home television set as a display device. Front End Processor - this acts as the device to control data communications from and to the consumer. The system is centralized with the FEP providing for both the data link control and network control functions. CPU - the major controlling element of each regional system. Those systems a:e to be configured so that a single headend can serve a single large metropolitan area. Gateway - this provides for the interface between the local network of the cable system and the supplier's data center. The gateway supports such data interfaces such as X.25 protocols. Disk Files - these store videotex data plus all other support data. The videotex is stored locally and as such represents a distributed data base. Video Control Unit (VCU) - this device provides CPU interface to the video disc units (VDU) which store all the full motion video. It also controls the video switch unit (VSU) that interfaces the VDUto a given cable channel. It provides the video packetizing and scheduling necessary for the system multiplpxing. Cable Plant - the FEP and VSU interface directly with the cable plant at the headend. This plant is a hierarchical plant and allows a space division multiplexing to be employed. Channels are allocated for downstream video, downstream data and upstream data. These hardware elements represent the.major functions that are to be performed. The second key architectural element is the software architecture. In this design, the following are the software architectural elements. Communications Management - supporting both FEP and Gateway. Message ~illnagement - handing control over all local transactions. Applications - providing data and videotex support in text generation and formatting. DBMS - handling control of all data bases. Operating Systems - performing all OS functions. Performance Monitoring System Management Figures 2 and 3 depict how these major software architectural elements interact for both local and remote transactions TECHNOLOGY There are many technological factors that will impact the deployment of this design. There are those key ones that have been developed: the multiple access choice, the home terminal design, and the video disc control units. MULTIPLE ACCESS There are several multiple access schemes that have been considered for the consumer market, but two are currently in use. The first is a polling scheme used in Warner's QUBE system, and the second is a Carrier Sensed Multiple Access/Collision Detection (CSMA/CD) scheme used in the Cox Indax system. The QUBE polling scheme uses an adaptive polling table approach. The lowest poll rate is for the entertainment user who is polled once per polling cycle (about 20 seconds). Interactive users are polled once every 4 seconds while inactive and then once every 0.5 seconds when active. It is this variable poll rate that gives rise to the adaptive approach. The polling table "adapts" to the activity of an interactive terminal. The polling system was used to avoid the large noise ingress that results when there are 300,000 or more taps on a cable. This is done through the use of devices called bridger gate controllers (BGC's) which are synchronized with the polling cycle to lock onto various positions of the network. This technology has been shown to work efficiently for over five years in a cable environment and over 20 years in standard computer environments. CSMA/CD works on the principle of waiting to transmit until the channel seems clear. If a collision occurs the device re-transmits. Polling is centralized; CSMA/CD is fully distributed. This latter technology was designed for baseband local area networks (one mile in length) and has problems in broadband cable networks (1,700 miles of cable). The following are critical factors in CSMA/CD. Long cable runs require low data rates and thus many data channels to handle total data requirements

5 Attempts to reduce round trip delay times result in increased plant hardware which is very costly. Home terminal costs are greatly increased since CSMA/CD requires more protocol intelligence. Considering that 80% of all total invested capital may be in home terminals, this has a significant economic effect. :L Figure 4 depicts a tradeoff analysis of polling versus CSMA/CD. Clearly the efficiency of polling (the ratio of effective channel throughput to data rate) is dramatically higher. As CSMA/CD data rate increases, both efficiency and capacity decrease and bandwidth increases. The conclusion that is drawn from this is that polling is clearly the more effective multiple access scheme for consumer networks. CSMA/CD is a technological elegance whose cost is not justified. HOME TERMINALS The home terminal (HT) must have the capabifity of supporting a host of service offerings and so in a "user-friendly" format. Consumer market studies seem to indicate that keyboards (full alphanumeric) rather than keypads (numbers only) are preferred. Cost is also a critical factor. A terminal target cost of $300 is considered acceptable. This ensures a three-year payback period at $10 per month allocated user fee. This terminal must provide the following key functions: CATV converter capability TV interface DLC interface to the cable Modem capability 8K of RAM for local storage SK to 10K ROM for basic OS and commands Keyboard interface Minimum 8 bit CPU capability Graphics capability Force tuning capability of automatically switch cable channels. VIDEO DISC SYSTEM The heart of this network is the video disc system (VDS) composed of the VCU, VSU and VDU. This system must be capable of accessing any segment of anyone of up to 400 VDU s. The VSU must be capable of selecting outgoing channels and switch the VDU. The VCU must provide for sequencing, packetizing and control. TRENDS AND ISSUES The driving force in the development of cable local loop ~apability is considered to come from the consumer side. It is a system concept that integrates video, voice and data into a new medium that meets a market need. It is positioned to compete with other successful marketing channels. It represents a set of interesting technical challenges as relates to plant, storage and home terminal. There are, however, several key trends and issues that still need to be resolved. Interconnect - cable is a very frag~ented transmission medium. How can the many MSO's interconnect their systems and when. Hybrid Designs - cable is still not ubiquitous. Telephone is. Hybrid designs integrating both are key. Two-way Systems - roughly only 10% of all cable plant is now two-way operational. At issue is interim use of hybrids until two-way is activated. Privacy and Security - privacy is a difficult concept. It assumes that people give only what they want in information and that it is safeguarded from unauthorized use. Legal attempts to ensure privacy have resulted in "watchers" to ensure it. But this begs the question of who "watches" the "watchers." Security, on the other hand, provides local data transmissions integrity and is ensured by encryption. At issue is how much encryption and at what cost. New Medium - what cable, a smart HT and distributed data base create is a new medium. What the consumer views over this medium will evolve in time. The challenge is to provide a clear understanding of the potential of this medium, its dimensions of responsivity from the user, that far surpasses anything that we have yet seen in communication networks. What we have presented in this paper is a new approach to facilitate the provision of information, the ability to perform transactions and the utilization of entertainment. This system has been developed at Warner Amex in response to market demands but represents a dramatic technological as well as sociological challenge. It removes the passive television behavior and adds the dimension of choice, selectivity and selfsegmentation. This represents a significant operational challenge. It also represents the key difference between text only and systems that have full motion video on demand

6 2SO ~lt,.eut '"... S~ftr0411'(rn.RErUee_, DIS'S Totel. >l 7> ) c.,.cltl' tot. _Uth "1'o111nl (000' 21 U. 156Dp. Opt D, auut.\oint,... Ca_ CllWCD "",eiv. FI<Ul 2: 9/!HTtAAcTlaos 'lil ljxal TR.tHSACTIlJI